Program

  • 08:00
  • 08:15
    • - Registration & Refreshments
    • - Welcome
    • - Prof. Shai Shalev-Shwarts,Mobileye
    • - Ayelet Avrahami,Elta
    • - Coffee Break & Visit the Exhibition
    • - Editing real images with pre-trained flow modelsProf. Tomer Michaeli,Technion - Israel Institue of Technology
    • - Prof. Iris Barshack,Sheba Medical Center & Ariel University
    • - Coffee Break & Visit the Exhibition

Close
  • 08:00
  • 08:10

  • Track 1: Generative AI

    • - OmnimatteZero: Fast Training-free Object Removal and Composition in Videos with Pre-trained Diffusion ModelsDvir Samuel,OriginAI & Bar-Ilan University
    • - Latent Space JAM: Layout-Guided Video GenerationMoshe Mandel & Asaf Joseph,Lightricks
    • - Consistent Amortized Clustering via Generative Flow NetworksIrit Chelly,Ben Gurion University of the Negev
    • - Black-box Adversarial Attack on Stable Diffusion ModelsShachar Shmueli,Ben Gurion University of the Negev
    • - Leveraging Diffusion Models towards PE Early Diagnosis using CXRsNoa Cahan,Tel Aviv University

  • Track 2: LLMs & Vision

    • - The Geometry and Likelihood Structure of CLIP EmbeddingsMeir Yosef Levi,Technion - Institute of Technology
    • - The Scientific Image Arms Race: How We Detect AI Generated Figures in the WildDori Gaton,Data Compass AI & Proofig AI
    • - CLIMP: Contrastive Language-Image Mamba PretrainingNimrod Shabtay,IBM Research & Tel Aviv University
    • - Ofir Bibi,Lightricks
    • - Amir Alush,Visual Layer

  • Track 3: 3D Vision

    • - Looking Into the Water by Unsupervised Learning of the Surface ShapeOri Lifschitz,University of Haifa
    • - Retrieval-Augmented 3D Vision: Analysis and Generation in the Long TailSagie Benaim,The Hebrew University of Jerusalem
    • - Multi-Modal Geologic Intelligence: 3D Inversion and Map Synthesis via Generative Foundation ModelsMichal Holtzman Gazit,Earth Dynamics
    • - Mesh Refinement from Multi-View RGB Using Image-Predicted Surface NormalsZvi Stein,Aligntech
    • - Extreme Rotations Estimation In The WildHani Bezalel,Tel Aviv University

Close
  • 08:00
  • 08:15
    • - Lunch Break

Close
  • 08:00
  • 08:12

  • Track 4: Autonomous Systems

    • - BINA: Bootstrapped Intelligence for Novel AdaptationOr Kozlovsky,Bluewhite Robotics
    • - BADAS: Context-Aware Ego-Centric Collision Prediction Using Real-World Dashcam DataRoni Goldshmidt,Nexar
    • - Seed-to-Seed: Unpaired Image Translation in Diffusion Seed SpaceOr Greenberg,General Motors

  • Track 5: Visual Surveillance & Object Tracking

    • - Object Detection and Tracking in Live Streams Using Textual and Visual Detailed DescriptionsTom Hirshberg,Microsoft
    • - Efficient Multi-Camera Multi-Person Tracking in Sparse CCTV Layouts for Loss Prevention in Retail SpacesGilad Erlich Trigo,Retail
    • - Data Efficiency Estimation from Tiny Datasets: Sparse Polarization Biometric Case StudyAbraham Pelz,Samsung

  • Track 6: Agents & RAGS

    • - Introduction to Multi-Agent Architecture PatternsRafael Ben Ari,General Motors
    • - When AI Agents Should Ask for Help - Building Reliable Human-AI SystemsOr Levi,Zefr
    • - Real-World Multi-Modal RAG: Innovative Benchmarking and Efficient Visual Document RetrievalAdi Raz,IBM

Close
  • 08:00
  • 08:15
    • - Visit the Exhibition & Poster Session

Close
  • 08:00
  • 08:10

  • Track 7: Advanced Learning Theory

    • - A label-efficient active learning framework for medical image segmentation: from cold start to increased scaleDevon Levy,Haifa University
    • - Adaptive Resolution Processing in Vision-Language ModelsDr. Eli Schwartz,IBM Research
    • - FastJAM: a Fast Joint Alignment Model for ImagesOmri Hirsch,Ben Gurion University of the Negev
    • - Machine Unlearning of Deep Neural Networks: The Effect of OverparameterizationGal Alon,Ben Gurion University of the Negev

  • Track 8: "Real-World" AI

    • - Affordable Thermal Imaging: Overcoming Accuracy and Resolution Limits with AIIftach Klapp,Volcani Institute
    • - Scaling Convolutional Neural Networks for Tabular Data via Correlation-Based Image TransformationsLoai AbdAllah,The Max Stern Yezreel Valley College
    • - Warp and Render: A Dual-Network Framework for Geometry-Controlled Simulation in Semiconductor Process DiagnosticsRan Itay,Applied Materials
    • - Unified Sports Perception: Single-Pass Extraction of Dense Player Metadata via Lightweight VLMsIgal Dmitriev and Ofir Liba,WSC-Sports

  • Track 9: AI in Healthcare & Scientific Discovery

    • - Advancing AI in Radiology Research with NVIDIA Clara Open Medical Imaging ModelsAmit Bleiweiss,NVIDIA
    • - The Micro-Movie of Cancer: Uncovering Therapy Resistance through High-Content Cell TrackingJeries khoury,Tel Aviv University
    • - Adapting Long-Video Masked Autoencoders to High-Speed Brain ImagingNatalya Segal,Bar Ilan University
    • - Automatic 3D Surface Reconstruction of the Left Atrium from Unorganized ContoursLorina Dascal,Abbott Laboratories

Close
  • 08:00
  • 08:15
    • - Happy Hour, Visit the Exhibition & Poster Session
    • - A Geometric-Probabilistic View of Diffusion and Manifold Projection for Image RestorationDr. Leah Bar,Autobrains & Tel-Aviv University
    • - Granite Vision – the VLM for Enterprise WorkflowsTal Drory,IBM Research
    • - Deep-learning for Geometric AlignmentProf. Oren Freifield,Ben Gurion University
    • - Idan Bassouk,Aidoc
    • - Prof. Ilan Tsarfaty,Tel Aviv University

Close